cvs2svn_lib/database.py

   1 # (Be in -*- python -*- mode.)
   2 #
   3 # ====================================================================
   4 # Copyright (c) 2000-2009 CollabNet.  All rights reserved.
   5 #
   6 # This software is licensed as described in the file COPYING, which
   7 # you should have received as part of this distribution.  The terms
   8 # are also available at http://subversion.tigris.org/license-1.html.
   9 # If newer versions of this license are posted there, you may use a
  10 # newer version instead, at your option.
  11 #
  12 # This software consists of voluntary contributions made by many
  13 # individuals.  For exact contribution history, see the revision
  14 # history and logs, available at http://cvs2svn.tigris.org/.
  15 # ====================================================================
  16
  17 """This module contains database facilities used by cvs2svn."""
  18
  19
  20 import sys
  21 import os
  22 import cPickle
  23
  24 from cvs2svn_lib.common import DB_OPEN_NEW
  25 from cvs2svn_lib.common import warning_prefix
  26 from cvs2svn_lib.common import error_prefix
  27 from cvs2svn_lib.log import logger
  28
  29
  30 # DBM module selection
  31
  32 # 1. If we have bsddb3, it is probably newer than bsddb.  Fake bsddb = bsddb3,
  33 #    so that the dbhash module used by anydbm will use bsddb3.
  34 try:
  35   import bsddb3
  36   sys.modules['bsddb'] = bsddb3
  37 except ImportError:
  38   pass
  39
  40 # 2. These DBM modules are not good for cvs2svn.
  41 import anydbm
  42 if anydbm._defaultmod.__name__ in ['dumbdbm', 'dbm']:
  43   logger.error(
  44       '%s: cvs2svn uses the anydbm package, which depends on lower level '
  45           'dbm\n'
  46       'libraries.  Your system has %s, with which cvs2svn is known to have\n'
  47       'problems.  To use cvs2svn, you must install a Python dbm library '
  48           'other than\n'
  49       'dumbdbm or dbm.  See '
  50           'http://python.org/doc/current/lib/module-anydbm.html\n'
  51       'for more information.\n'
  52       % (error_prefix, anydbm._defaultmod.__name__,)
  53       )
  54   sys.exit(1)
  55
  56 # 3. If we are using the old bsddb185 module, then try prefer gdbm instead.
  57 #    Unfortunately, gdbm appears not to be trouble free, either.
  58 if hasattr(anydbm._defaultmod, 'bsddb') \
  59     and not hasattr(anydbm._defaultmod.bsddb, '__version__'):
  60   try:
  61     gdbm = __import__('gdbm')
  62   except ImportError:
  63     logger.warn(
  64         '%s: The version of the bsddb module found on your computer '
  65             'has been\n'
  66         'reported to malfunction on some datasets, causing KeyError '
  67             'exceptions.\n'
  68         % (warning_prefix,)
  69         )
  70   else:
  71     anydbm._defaultmod = gdbm
  72
  73
  74 class Database:
  75   """A database that uses a Serializer to store objects of a certain type.
  76
  77   The serializer is stored in the database under the key
  78   self.serializer_key.  (This implies that self.serializer_key may not
  79   be used as a key for normal entries.)
  80
  81   The backing database is an anydbm-based DBM.
  82
  83   """
  84
  85   serializer_key = '_.%$1\t;_ '
  86
  87   def __init__(self, filename, mode, serializer=None):
  88     """Constructor.
  89
  90     The database stores its Serializer, so none needs to be supplied
  91     when opening an existing database."""
  92
  93     # pybsddb3 has a bug which prevents it from working with
  94     # Berkeley DB 4.2 if you open the db with 'n' ("new").  This
  95     # causes the DB_TRUNCATE flag to be passed, which is disallowed
  96     # for databases protected by lock and transaction support
  97     # (bsddb databases use locking from bsddb version 4.2.4 onwards).
  98     #
  99     # Therefore, manually perform the removal (we can do this, because
 100     # we know that for bsddb - but *not* anydbm in general - the database
 101     # consists of one file with the name we specify, rather than several
 102     # based on that name).
 103     if mode == DB_OPEN_NEW and anydbm._defaultmod.__name__ == 'dbhash':
 104       if os.path.isfile(filename):
 105         os.unlink(filename)
 106       self.db = anydbm.open(filename, 'c')
 107     else:
 108       self.db = anydbm.open(filename, mode)
 109
 110     # Import implementations for many mapping interface methods.
 111     for meth_name in ('__delitem__',
 112         '__iter__', 'has_key', '__contains__', 'iterkeys', 'clear'):
 113       meth_ref = getattr(self.db, meth_name, None)
 114       if meth_ref:
 115         setattr(self, meth_name, meth_ref)
 116
 117     if mode == DB_OPEN_NEW:
 118       self.serializer = serializer
 119       self.db[self.serializer_key] = cPickle.dumps(self.serializer)
 120     else:
 121       self.serializer = cPickle.loads(self.db[self.serializer_key])
 122
 123   def __getitem__(self, key):
 124     return self.serializer.loads(self.db[key])
 125
 126   def __setitem__(self, key, value):
 127     self.db[key] = self.serializer.dumps(value)
 128
 129   def __delitem__(self, key):
 130     # gdbm defines a __delitem__ method, but it cannot be assigned.  So
 131     # this method provides a fallback definition via explicit delegation:
 132     del self.db[key]
 133
 134   def keys(self):
 135     retval = self.db.keys()
 136     retval.remove(self.serializer_key)
 137     return retval
 138
 139   def __iter__(self):
 140     for key in self.keys():
 141       yield key
 142
 143   def has_key(self, key):
 144     try:
 145       self.db[key]
 146       return True
 147     except KeyError:
 148       return False
 149
 150   def __contains__(self, key):
 151     return self.has_key(key)
 152
 153   def iterkeys(self):
 154     return self.__iter__()
 155
 156   def clear(self):
 157     for key in self.keys():
 158       del self[key]
 159
 160   def items(self):
 161     return [(key, self[key],) for key in self.keys()]
 162
 163   def values(self):
 164     return [self[key] for key in self.keys()]
 165
 166   def get(self, key, default=None):
 167     try:
 168       return self[key]
 169     except KeyError:
 170       return default
 171
 172   def close(self):
 173     self.db.close()
 174     self.db = None
 175
 176